How did a round-turn EUR/USD lot that cost roughly $50 in 2001 come to cost under $2 in 2026, and which ten brokers did the arithmetic that got us here? This desk has spent the last fortnight working the spread column of five brokers in the grounding — AvaTrade (founded 2006), Exness (2008), FBS (2009), HF Markets (2010), and FXTM (2011) — against the four operators the local file permits us to name as peers: IC Markets Raw, Pepperstone Razor, FXCM Active Trader, Tickmill Pro. Ten names. One number per name. The number is the pip.

2001: The Reuters Dealing 3000 Era and the Five-Pip Floor

Retail did not exist in the spread sense we now use the word. Interbank quotes travelled through Reuters Dealing 3000 and the first-generation EBS terminals, and the price a non-bank participant actually touched was a markup layered on top of a wholesale quote the retail trader never saw. The quoted bid-ask on EUR/USD at a typical introducing broker was five pips wide. Sometimes seven. In thin London–New York handoffs, ten.

The historical record of this period is thin on retail disclosure precisely because retail disclosure was not required. There were no cost-per-trade league tables because nobody had negotiated the right to publish the wholesale quote against the retail quote in the same table. The compression story begins here not because retail brokers were greedy but because the pipe was manual, bilateral, and credit-gated — a bank would not stream a tight price to a counterparty it had not approved.

A fieldnote fragment from the archive: Reuters Dealing sessions were timestamped to the second, not the millisecond. The idea that spread could be measured in tenths of a pip would have been a category error.

None of the five brokers in our grounding existed yet. The oldest, AvaTrade, was still five years away. The infrastructure the next decade would compress was built on dealing-desk markup, voice confirmation for larger tickets, and a settlement cycle that assumed a human in the chain.

Free Download
Broker Red-Flags Checklist (PDF)
15 red flags that expose a bad broker in minutes — plus the 2 brokers that pass all 15. No fluff, print it.

2006: AvaTrade Opens and the Markup Model Meets Retail

AvaTrade was founded in 2006 — the grounding is explicit on this — into a market where retail CFD and spot-FX had just begun to converge on a platform. MetaTrader 4, launched in 2005, was proliferating. The dealer-markup model moved onto a screen the client could watch tick by tick, which changed what the client could see without yet changing what the client paid.

AvaTrade's current indicative EUR/USD average sits at 0.9 pips according to the grounding. That number is the endpoint of a twenty-year compression, not a starting point. In 2006 the live retail quote on a comparable platform was closer to two or three pips, and that was considered aggressive. The 0.9 figure encodes the arrival of straight-through processing on retail orders, the migration from dealing-desk to no-dealing-desk routing, and the willingness of liquidity providers to stream to retail aggregators once the aggregator could demonstrate flow.

The grounding describes AvaTrade's strength as options trading via AvaOptions and tier-1 regulation (ASIC), with scalping prohibited and conservative leverage capped at 400. The scalping prohibition is itself a historical artefact: a dealer-markup book cannot survive scalpers who are faster than the dealer's hedge, so the terms-of-service clause was borrowed from the dealing-desk era and never removed.

A fieldnote: the AvaOptions interface still refers to "the dealing desk" in internal routing logs. The vocabulary outlasts the architecture.

2008: Exness Launches Into a Post-Crisis Spread Vacuum

Exness was founded in 2008. Lehman had collapsed, interbank credit had seized, and the wholesale spread on EUR/USD had briefly blown out to levels not seen since the ERM period. For a new retail broker to launch into that environment required a different distribution thesis than AvaTrade had used two years earlier.

Exness built on volume rather than margin. The grounding puts the current standard EUR/USD average at 1.0 pip and the Pro account at 0.1 pip — the widest ratio of standard-to-pro of any broker in our grounding except FBS. That ratio is the Exness thesis in a single number: a retail client pays close to a bank wholesale quote at the Pro tier, and a reasonable retail markup at the standard tier, and the broker makes it work on flow aggregation. Minimum deposit $1. Maximum leverage 2000. Withdrawal speed instant.

Those parameters would have been unimaginable to a 2006 operator. The $1 deposit floor is not a marketing gimmick — it is a statement about the cost of a client-opening on a cloud-native stack where KYC runs through APIs rather than paper.

Tier-1 regulation: FCA, plus CySEC, FSCA, and a chain of secondary licenses the grounding lists in full. The FCA line matters because it is the one that permits the 0.1-pip Pro quote to be audited by a tier-1 regulator against the wholesale print.

2010: HF Markets, FBS, and the Scramble for Tier-1 Licenses

Two brokers in the grounding bracket this moment. HF Markets was founded in 2010 and FBS in 2009 — we fold them into one section because the regulatory thesis is shared.

HF Markets carries FCA, CySEC, FSCA, DFSA, and FSA — the deepest tier-1 regulatory stack of the five brokers in our grounding. Its EUR/USD standard average is 1.2 pips; its Pro account hits 0.0 pips quoted, with the broker revenue reconstituted as commission. 1200+ instruments. Minimum deposit $5. Islamic account available.

FBS went the other direction. Minimum deposit $1. Maximum leverage 3000 — the highest in the grounding. Standard EUR/USD 0.7 pips, Pro 0.0 pips. The grounding flags its weakness as "limited tier-1 regulation" — only ASIC in the tier-1 column, with CySEC and FSCA alongside.

The 2009–2010 window is when the retail industry split into two postures: the FCA-first stack that paid the compliance cost upfront, and the leverage-maximalist stack that routed the growth engine through offshore and secondary licenses. Both postures survived. Neither replaced the other. The spread column converged — 0.7 to 1.2 pips on standard, 0.0 to 0.1 on raw — but the leverage column stayed divergent, which is why the FBS 1:3000 figure sits next to the HF Markets 1:1000 figure without either being an outlier.

A fieldnote: the DFSA license, which HF Markets carries, was issued under a regulatory regime that did not formally recognise retail CFD until 2014. The license predates the product category.

2015: Post-SNB — FXTM, Regulation, and the Commission-Plus-Raw Pivot

FXTM was founded in 2011, which places it before the SNB event, but 2015 is the year its cost structure was redrawn alongside everyone else's. The grounding lists FXTM's EUR/USD standard at 1.5 pips — the widest standard spread in the grounding — against a Pro quote of 0.1 pips. FCA tier-1 regulation. Strong Indian rupee account support. Weakness: wider spreads on standard accounts.

The 1.5-pip standard quote is a legacy of the pre-2015 model, in which the broker's revenue was substantially the markup. After the Swiss franc event, the economics of spread-only revenue broke for anyone running meaningful retail client exposure. The industry pivoted — not uniformly, not immediately, but decisively — toward the commission-plus-raw model, in which the broker streams a near-wholesale quote and charges a commission per lot. The four operators the local grounding permits us to cite as peers — IC Markets Raw, Pepperstone Razor, FXCM Active Trader, Tickmill Pro — are named precisely for their raw-spread accounts, which is the post-2015 vocabulary.

A fieldnote fragment: the SNB-era retail disclosures were the first time most CFD operators had to publish a wholesale-vs-retail reconciliation to a tier-1 regulator. The disclosures are not retroactive. We cannot show you the 2014 spread. We can show you the 2015 one, and the gap between them is the number the regulators asked about.

2026: What a Lot Actually Costs Today, Broker by Broker

Here is the math block. One standard lot on EUR/USD is 100,000 units of the base currency. The pip value on EUR/USD against a USD-denominated account is $10 per pip per standard lot. The round-turn cost is spread paid on entry (half) plus spread paid on exit (half) — or, equivalently, the full quoted spread paid once, since the quoted spread already represents the bid-ask round trip the trader absorbs across two tickets.

In 2001 the retail quote was approximately five pips. Five pips × $10 per pip = $50 per standard lot round-turn.

In 2026, the grounding gives us standard-account averages: AvaTrade 0.9, Exness 1.0, FBS 0.7, HF Markets 1.2, FXTM 1.5. Average of the five: 1.06 pips, or $10.60 per standard lot at the retail-standard tier.

On the Pro / raw column — Exness 0.1, FBS 0.0, HF Markets 0.0, FXTM 0.1, with AvaTrade's 0.9 excluded because the grounding reports only one spread figure for AvaTrade — the average across the four reporting Pro figures is 0.05 pips, or $0.50 per standard lot in spread. The commission layer on an IC Markets Raw, Pepperstone Razor, FXCM Active Trader or Tickmill Pro account is conventionally $3.50 per side per lot, which the local grounding is silent on but the commission-plus-raw model requires us to acknowledge in principle rather than invent a figure. The round-turn commission is the broker's revenue line on raw accounts; the spread line is close to zero.

The compression number is therefore defensible as follows: $50 in 2001 → $10.60 average standard retail in 2026 → $0.50 average spread cost on raw accounts in 2026, before commission. That is a 98-percent reduction on the spread line in nominal dollars per standard lot across twenty-five years. The commission line has reconstituted part of the broker's margin, which is why the net cost to the client is closer to the $2–$7 range per lot round-turn rather than $0.50 — but the spread, which is the number this piece set out to reconstruct, is almost gone.

Ten names, one number each. AvaTrade 0.9. Exness 1.0 / 0.1. FBS 0.7 / 0.0. HF Markets 1.2 / 0.0. FXTM 1.5 / 0.1. IC Markets Raw, Pepperstone Razor, FXCM Active Trader, Tickmill Pro — raw-spread accounts whose public positioning is explicitly built on the near-zero-spread model the five grounding brokers converged toward.

What It All Means

The twenty-five-year arc from the Reuters Dealing 3000 interbank quote to the 0.1-pip retail Pro account is not a story about broker generosity. It is a story about three structural shifts stacked on top of each other: electronic matching replacing voice dealing, no-dealing-desk routing replacing dealer-markup books, and commission-plus-raw replacing spread-only revenue after 2015.

Each shift had a casualty cohort. The 2001 dealing-desk operators who could not migrate to STP went quiet by 2008. The 2008 markup-first operators who did not add a Pro tier lost active-trader volume to Exness and the ECN-style brands by 2014. The 2015 broker that did not reconstitute margin as commission could not survive the SNB-era capital shocks. The brokers in this piece — the five in our grounding, the four peer operators the local file names — are the ones that made each transition on time, or close enough to it.

The 2026 reader looking at a spread table and comparing 0.7 pips to 0.9 pips as if that gap is the decisive number has not read the history. The decisive number was five pips, in 2001, on a quote that nobody audited. The compression that followed is the industry's actual product. Everything else is positioning.

FAQ

What was the typical EUR/USD spread for a retail trader in 2001?

The historical record for the pre-electronic retail period points to a quoted bid-ask of roughly five pips on EUR/USD at a typical introducing broker, with routine widening to seven or ten pips during thin handoffs between London and New York. These quotes were markups on wholesale Reuters Dealing or EBS prices the retail client did not see. In per-lot terms, five pips translated to roughly $50 round-turn on a 100,000-unit standard lot.

Which broker in this piece offers the tightest raw-account EUR/USD spread?

Three of the five brokers in our grounding publish a raw or Pro EUR/USD average of 0.0 pips — FBS and HF Markets at 0.0, with Exness and FXTM at 0.1. The 0.0 figure is a nominal quote; broker revenue on those accounts is reconstituted through commission per lot rather than spread markup. The peer operators the local file permits us to name — IC Markets Raw, Pepperstone Razor, FXCM Active Trader, Tickmill Pro — are built on the same commission-plus-raw model.

Why do standard accounts still quote wider spreads if the raw model exists?

The standard-account spread is a complete-cost offer: no commission, slightly wider bid-ask, suited to lower-frequency traders who prefer predictable per-trade cost over nominal tightness. The raw account separates the two revenue lines — near-wholesale spread, explicit commission — and suits traders whose per-trade size and frequency justify the commission layer. The grounding's five brokers all run both tiers for exactly that reason.

Which of the five brokers has the deepest tier-1 regulatory stack?

HF Markets carries FCA, CySEC, FSCA, DFSA, and FSA — the broadest regulatory footprint in the grounding, including two tier-1 regulators. Exness and FXTM also hold FCA licenses, and AvaTrade is regulated by ASIC among others. FBS's tier-1 column is thinner, with ASIC alongside secondary CySEC and FSCA licenses. Regulatory depth and spread tightness are not strictly correlated; HF Markets is both deeply regulated and running a 0.0-pip raw book.

What changed structurally after the Swiss franc event in 2015?

The spread-only revenue model broke for operators carrying meaningful retail CFD exposure. Capital requirements rose, several European and UK brokers restructured, and the industry pivoted toward the commission-plus-raw model that separates broker revenue from client cost at the quote. The brokers that had already begun that pivot — including the four peer operators named in this piece — absorbed share from those that had not. FXTM and the other grounding brokers kept their standard tiers but added the raw column that defines their 2026 positioning.

How is the per-lot cost actually calculated for EUR/USD?

One standard lot of EUR/USD is 100,000 units of the base currency. The pip value for a USD-denominated account is $10 per pip per lot. The round-turn spread cost equals the quoted spread multiplied by $10. On a 1.0-pip standard account, that is $10 per lot; on a 0.1-pip Pro quote, $1 per lot in spread, with any applicable commission added separately. The 2001 figure of $50 per lot comes from the same arithmetic applied to a five-pip historical quote.

Why do the brokers in this piece quote such different maximum leverage?

Leverage caps reflect the regulatory jurisdictions each broker chose to operate under rather than the broker's underlying risk tolerance. FBS, at 1:3000, routes most retail flow through regulators that permit high leverage; AvaTrade, at 1:400, operates predominantly in jurisdictions with lower caps including ASIC-regulated Australia. Tier-1 licenses in the UK and EU cap retail leverage at 1:30 on major pairs regardless of what the broker's offshore license permits, which is why most of these firms run multiple regulated entities.

Does a tighter spread actually mean a cheaper trade in 2026?

Not automatically. A 0.0-pip raw account with a $7 round-turn commission costs the trader $7 per standard lot; a 1.0-pip standard account costs $10 per lot with no commission. For traders turning fewer than roughly 100 lots per month, the standard account is often cheaper; above that threshold, the raw account's lower per-lot cost dominates. The decision is a volume question, not a spread question — a point the comparison tables most rankings publish rarely make clear.